Composite Structures 94 (2012) 2285–2295

Contents lists available at SciVerse ScienceDirect

Composite Structures
journal homepage: www.elsevier.com/locate/compstruct

Flexible cellular solid spokes of a non-pneumatic tire

Jaehyung Ju a, Doo-Man Kim b,⇑, Kwangwon Kim b
a
Department of Mechanical and Energy Engineering, University of North Texas, 1155 Union Circle #311098, Denton, TX 76203-5017, USA
b
School of Aerospace and Mechanical Engineering, Korea Aerospace University, Goyang-City, Gyeonggi, Republic of Korea

a r t i c l e i n f o a b s t r a c t

Article history: Non-pneumatic tires (NPTs) have been introduced with a compliant cellular solid spoke component
Available online 2 January 2012 which functions as the air of the pneumatic tire. In this paper, hexagonal honeycomb spokes for a high
fatigue resistance design are investigated by seeking compliant hexagonal structures that have low local
Keywords: stresses under macroscopic uni-axial loading. Using the honeycomb mechanics, two cases of hexagonal
Cellular solids honeycombs are designed: (i) the same cell wall thickness and (ii) the same load carrying capacity.
Lattice structures The elastic limits of the hexagonal honeycombs are obtained from the ABAQUS finite element code con-
Honeycombs
sidering the geometric nonlinearity of a cellular structure associated with the cell wall buckling and
Compliant design
Non-pneumatic tire (NPT)
bending. The compliant cellular structures having low local stress values are applied to the honeycomb
Airless tire spokes of an NPT for the structural validation and the local stresses of the honeycomb spokes are inves-
tigated under the same vertical loading conditions. Hexagonal honeycombs with a highly positive cell
angle have low local stresses and low mass under the same vertical load carrying capability; the Type
C honeycomb spokes in this study.
Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Two dimensional prismatic cellular materials of periodic

Since the release of the first commercial pneumatic bicycle tire
by Dunlop in 1888, the pneumatic tire has been dominant in the
world tire market for more than 100 years due to four major
advantages it has over a rigid wheel: (i) low energy loss on rough
surfaces, (ii) low vertical stiffness, (iii) low contact pressure, and
(iv) low mass [1]. However, the pneumatic tire has several disad-
vantages as well: (i) the possibility of catastrophic damage – a flat
while driving, (ii) the required maintenance for proper internal air
pressure, and (iii) the complicated manufacturing procedure [1].
From the early 1920s, efforts were made to develop a resilient
wheel by adding wire spokes in order to deliver a quality pneu-
matic tire while guaranteeing safety [2]. Since then, several engi-
neers have also attempted to develop non-pneumatic tires (NPTs)
by filling an elastomer or by building polygon-like spokes to re-
place the air of a pneumatic tire [3–15]. There have been recent
innovations with respect to airless tires; NPTs have emerged con-
sisting of flexible polygon spokes and an elastomer layer having in-
ner and outer rings [12–14]. Considering the NPT structure, the
spokes undergo tension–compression cyclic loading while the tire
rolls [12–14]. Therefore, it is important to minimize the local stres-
ses of spokes when under cyclic loading while driving. In other
words, fatigue resistant spoke design takes on greater importance.
⇑ Corresponding author.
E-mail address: dmkim@kau.ac.kr (D.-M. Kim).
microstructures are called honeycombs. Honeycombs have been
primarily used in lightweight sandwich structures for which a high
out-of-plane stiffness is desired [16,17]. In contrast to the highly
stiff and strong properties of the out-of-plane direction (the longi-
tudinal cell axes) associated with cell wall axial stiffness, the in-
plane properties are two to three orders of magnitude weaker than
those of the out-of-plane loading. For this reason, the mechanical
properties which exist under in-plane loading have been thought
to be the most limiting in terms of design applications. However, re-
cently, there have been efforts to use lower in-plane stiffness for
designing flexible honeycomb structures to be used in applications
that require high deformation under targeted loads. Examples of
such applications are flexible micro-electro-mechanical-system
(MEMS) structures [18] and aircraft morphing structures [19–22].
Hexagonal honeycombs are cellular materials employed in
various applications, and particulary in the design of light weight
structures. Since the initial investigations conducted by engineers
on the effective moduli and effective Poisson’s ratios of hexagonal
honeycombs based on the bending dominant structural property,
of cell walls some more refined models have also been suggested
[23–28]. Recently, in-plane effective moduli, yield strengths, and
buckling strengths with different cell types – square, hexagonal,
triangle, mixed squares and triangles, and diamond – have been
investigated [29–35]. Triangular, Kagome, and diamond cell hon-
eycombs are known to be extension dominant cell structures,
which are good for high modulus structural designs [29]. On the
other hand, square and hexagonal cell honeycombs are known to

0263-8223/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compstruct.2011.12.022
2286 J. Ju et al. / Composite Structures 94 (2012) 2285–2295
be bending dominant structures, which are good for flexible struc-
tural designs [31,32]. Hexagonal cell structures are known to be
flexible in both axial and shear loadings [33–35]. Moreover, hexag-
onal honeycombs can easily be tailored to have targeted in-plane
properties by changing the cell angle, the cell wall thickness, and
the cell length. Therefore, some hexagonal structures have the po-
tential for compliant structural design.
The spokes of an NPT are required to have both stiffness and
resilience under cyclic tension–compression loading. In general,
stiffness and resilience are conflicting requirements if a material
has a high modulus, it shows a low elastic strain limit, and vice ver-
sa. The challenge, then, is to design materials that have both high
stiffness and high resilience. Because there are no conventional
materials to satisfy the desired property, the solution must be
sought for through the huge design space afforded by cellular
materials base materials and cellular geometries.
Finite element (FE) analysis has been utilized extensively in the
simulation of tire models due to its capability to solve complicated
structural behaviors combining the nonlinearity of a material and
geometry [36–38]. In this study, a commercial FE code, ABAQUS
is used for a numerical experiment with NPTs having hexagonal
honeycomb spokes. The honeycomb spokes made of polyurethane
Fig. 1. Geometric parameters of hexagonal honeycombs and boundary conditions for tensile and compressive loadings.
(PU) are suggested as the key component of NPTs which ade-
quately replace air-filled pneumatic tires. In Section II, uni-axial
flexible hexagonal structures are introduced as the spokes of the
NPT. Section III introduces the NPT structures and materials used.
The structural performance of the designed honeycomb spokes
for a fatigue resistant design is discussed in Section IV.
2. In-plane flexibility of honeycombs subjected to uni-axial
loading
Flexible honeycomb spokes along the radial direction of an NPT
function as the air of a pneumatic tire. When designing flexible
spokes, the structures should have both load carrying capabilities
and resilience. The former property is identified with the modulus
and the latter with the elastic strain limit. Unfortunately, those
properties are the conflicting requirements in general materials;
in other words, materials having a high modulus show a low yield
strain. However, cellular solids have the potential to free them-
selves from these conflicting requirements when designing flexible
structures because the cell geometry can be controlled. Consider-
ing the required cyclic loading of honeycomb spokes when the
NPT rolls, a cellular geometry having lower local stresses is
 J. Ju et al. / Composite Structures 94 (2012) 2285–2295 2287

Fig. 2. Uni-axial elastic stress–strain behaviors of 7075-T6 hexagonal honeycombs when they are designed to be a E11 of 5 MPa.

preferred in terms of fatigue resistant cellular structural design. In deformed shape at the maximum elastic limit, and the correspond-
this section, the resilience of the hexagonal honeycombs is investi- ing dimensions of the cell geometries are shown in Fig. 2. It should
gated with varying cell geometries. be noted that the initial slopes of the stress–strain curves for all
Hexagonal honeycombs are generated with the cell wall thick- honeycombs are the same (5 MPa) as can be seen in Fig. 2.
ness, t, the vertical cell length, h, the inclined cell length, l, and Also shown in Fig. 2, the effective stress–strain curves of the
the cell angle, h, as illustrated in Fig. 1. In order to find a flexible honeycombs vary depending on the cellular geometry even though
hexagonal honeycomb structure in the uni-axial direction, a simple their effective uni-axial moduli are designed to be the same. A high
FE based numerical experiment with honeycombs is conducted. cell angle, h, causes cellular structures to have flexibility under uni-
When a tensile or compressive displacement loading, d, is applied axial loading; e.g., the honeycomb with h = 50o shows a high effec-
on the nodes of the right edge, the reaction force, F1, on the nodes tive tensile strain (6.2%) and compressive strain (6.6%). On the
of the edge are obtained. Nodes on the bottom edge are free to
move in the x1 direction and are fixed in the x2 direction; u2 = 0
on x2 = 0. On the other hand, nodes on the left edge are free to
move in the x2 direction and are fixed in the x1 direction; u1 = 0
on x1 = 0. Effective axial stress and strain in the x1 direction, which
are denoted as r11 and 11 , respectively, are given by
R1 R1 d1
r11 ¼ ¼ and e11 ¼ ð1Þ
A1 wL2 L1
where R1 is the total reaction forces of the nodes on the right edge
and A1 is the cross-sectional area perpendicular to the x1 direction.
l1 and l2 are the lengths of the honeycomb structures along the x1
and x2 directions, respectively. w is the width of the honeycomb
structures in the x3 direction.
Fig. 2 shows the uni-axial stress–strain behaviors of the hexag-
onal honeycomb structures with varying honeycomb geometries
when they are designed to have the same effective uni-axial mod-
ulus, e.g., a E11 of 5 MPa with an aluminum alloy, 7075-T6
(E = 72 GPa and m = 0.33). Each cell wall is modeled with a shear
deformable beam element (B22 in ABAQUS). The stress–strain
curves of the honeycombs for h = 10, 10, 30, and 50°, their Fig. 3. A NPT consisting of honeycomb spokes.
2288 J. Ju et al. / Composite Structures 94 (2012) 2285–2295
(a) Polyurethane for spokes
Fig. 4. Stress–strain behaviors of polyurethane [39] and rubber [40].
Fig. 5. Geometric parameters of honeycombs in the polar coordinate.
other hand, honeycombs having a low magnitude of cell angle do
not have much flexibility under uni-axial loading; e.g., the honey-
combs with h = 10° and 10° show low effective tensile strains
(1.2%) and compressive strains (5.8%) within the base material’s
elastic range; the yield strength of 7075-T6, rys is 500 MPa.
The primary loading mode of cellular structures is known to be
the bending of cell walls [23,28]. Observing the stress distribution
and the deformed shapes of honeycombs under uni-axial loading
as shown in Fig. 1, the inclined cell length, l, experiences a high
bending stress under the global uni-axial loading of honeycombs
in the x1 direction. According to the beam deflection theory, a
longer beam provides more deflection by a lateral load than a
shorter one when the beams have the same cross-section and are
made of the same material. Therefore, a high length of l of a
honeycomb provides more flexibility under uni-axial loading. At
a given effective honeycomb volume, a higher l/X1 shows a higher
elastic limit when the effective moduli of honeycombs are the
same (Figs. 1 and 2): l/X1 values are 1.02, 1.02, 1.15, and 1.56 for
h = 10, 10, 30, and 50°, respectively.
3. Honeycomb spoke tires and constituent materials
Fig. 3 shows an NPT made up of the honeycomb spokes and
tread components. The inner face-sheet is made of an aluminum
4.5
4
3.5
3
stress (MPa)
2.5
2
Tread
1.5 (rubber)
1
0.5
0
0% 10% 20% 30% 40% 50%
strain
(b) Rubber for tread
alloy (7075-T6; density, q = 2800 kg/m3, Modulus, E = 72 GPa, and
Poisson’s ratio, m = 0.33) and functions as a rigid hub. A 1 mm thick
aluminum alloy is used for the inner face sheet (hub). The outer
face-sheet, made of a high strength steel (ANSI 4340;
q = 7800 kg/m3, E = 210 GPa, and m = 0.29), is set to be 0.5 mm. This
enforces the tread rubber to be deformed by shear. Without the
outer face-sheet, the edges of the spokes over the contact zone
with the ground would buckle and cause an undesirable nonlinear
effect of the honeycombs. All interfaces, e.g., (i) the hub and the
spokes, (ii) the spokes and the outer-face sheet, and (iii) the out-
er-face sheet and the tread, are modeled with the tie constraint
in ABAQUS.
Polyurethane (q = 1200 kg/m3, E  32 MPa, shear modulus,
G  10.81 MPa, and m = 0.49) is used as the constituent material
of the honeycomb spokes (core) and its uni-axial, bi-axial, and
planar tension test results are available in the literature are
shown in Fig. 4a [39]. The three test results are used as a
hyperelastic material input of the polyurethane honeycomb
spokes in the FE ABAQUS code. The honeycomb spokes have a
5 mm wall thickness. The tread component is made of a rubber
(q = 1043 kg/m3, E  11.9 MPa, G  4 MPa, and m  0.49), whose
stress–strain behavior is also available in the literature, and is
shown in Fig. 4b [40]. It is also used for the hyperelastic
model in the FE ABAQUS code. The thickness of the tread is
 J. Ju et al. / Composite Structures 94 (2012) 2285–2295 2289

 3
set to be 15 mm. A shear deformable quadratic beam element t cos h
(B22 in ABAQUS) is used for the spokes, and the inner and Er ¼ ES ð2Þ
l ðh=l þ sin hÞ  sin2 h
outer face sheets. A quadratic plane strain element (CPE8R in
ABAQUS) is used to model the tread for FE analysis. The 2D  3
t ðh=l þ sin hÞ
plane strain element was chosen considering the overall low EX ¼ ES ð3Þ
deformation of NPTs in the lateral direction (the z direction in
l cos3 h
Fig. 3).  3
t ðh=l þ sin hÞ
GrX ¼ Es ð4Þ
l ðh=lÞ2 ð1 þ 2h=lÞ cos h
4. Design of honeycomb spokes
where Er , EX , and ErX are the effective moduli in the radial, the cir-
cumferential, and the shear directions, respectively. h, l, and h are
4.1. Effective properties of honeycomb spokes
the average values of the vertical cell lengths, the inclined cell
lengths, and the cell angles, respectively (Fig. 5). Es is the Young’s
Effective in-plane moduli of hexagonal honeycombs were
modulus of a base material and t is the cell wall thickness. The den-
developed by earlier honeycomb engineers using the beam theory,
sity of a honeycomb, q⁄, from CMT [22–24], is also given by
and these developments are collectively called cellular materials
theory (CMT) [23–25]. According to CMT, the effective moduli are t=lðh=l þ 2Þ
given by [24]
q ¼ qs ð5Þ
2 cos hðh=l þ sin hÞ

Fig. 6. Suggested honeycomb spoke design.

Fig. 7. Convergence study of tread mesh.

2290 J. Ju et al. / Composite Structures 94 (2012) 2285–2295

Fig. 8. Convergence study of spoke mesh.

where qs is the density of the constituent material. The relative den-

Table 1
Number of nodes and elements for the spokes and the tread. sity of the honeycomb, q⁄/qs, is often considered to characterize
honeycomb behaviors, which will be discussed later in this section.
Type Spokes Tread
Six types of honeycomb spokes are considered for NPTs as
Node Element Node Element shown in Fig. 6. Both regular honeycombs and auxetic honeycombs
Type A 47,101 47,196 2035 1628 having a negative effective Poisson’s ratio are used for the cellular
Type B 47,331 47,426 2035 1628 spoke design. As shown in Section II, the ratio of the inclined cell
Type C 49,686 49,781 2035 1628
length to the cell height, l/X1 is the critical factor to design in-plane
Type D 51,148 51,243 2035 1628
Type E 56,183 56,278 2035 1628
flexible structures for simple tension–compression loading. There-
Type F 65,264 65,359 2035 1628 fore, the cell angle may not be important when designing the in-
plane flexible structures. In other words, the in-plane flexure

Fig. 9. Deformed shapes of NPTs under a vertical displacement, d(=20 mm) at the center.
 J. Ju et al. / Composite Structures 94 (2012) 2285–2295 2291

Table 2 ness and (ii) the same load carrying capability. A 2D NPT model for
Dimension of honeycomb spokes and effective properties when designed to be the a FE based numerical analysis with ABAQUS/standard will be pre-
same cell wall thickness, t.
sented in Sections 4.2–4.4 for the flexible spoke design. A 3D FE
l (mm) h (mm) h (degree) t (mm) q⁄/qs Er/Es EX/Es GrX/Es model may be needed for the accurate characterization of an NPT
Type A 26.25 36.66 15.76 5.00 0.20 0.054 0.013 0.0016 deformation. However, when considering a higher modulus and a
Type B 29.65 28.52 31.50 5.00 0.20 0.010 0.011 0.0031 lower deformation in the lateral direction (the z direction in
Type C 37.21 16.74 47.14 5.00 0.20 0.003 0.009 0.0110 Fig. 5), a 2D NPT model appears to be sufficient for use in investi-
Type D 26.25 50.93 15.76 5.00 0.23 0.054 0.013 0.0007
Type E 29.65 59.43 31.50 5.00 0.27 0.010 0.011 0.0004
gating the tire performance from a design perspective.
Type F 37.21 71.14 47.14 5.00 0.33 0.003 0.009 0.0002
4.2. Mesh convergence study of tread and spokes

behavior of auxetic honeycombs may not be greatly different from Optimum numbers of elements (mesh size) for the tread and
that of regular honeycombs. However, taking into account addi- the spokes are found in this section. The element numbers of the
tional boundary conditions – bending at the contact edges and tread and the spokes are controlled with the global mesh size op-
transverse compression near the contact patch, the cell angle tion in ABAQUS. For a mesh convergence study of the tread, when a
may affect the macroscopic spokes’ behavior in the NPT design. displacement load is applied, reaction force at the hub center and
In this paper, two different spoke design concepts are suggested maximum local stresses are checked for regular and auxetic honey-
with the six honeycomb geometries – (i) the same cell wall thick- combs when a fixed global mesh size of spokes (0.0061 in ABAQUS)

500 500

reaction force at the center (N)

reaction force at the center (N)

450 450 Type D

Type A
400 400 Type E
Type B
350 350 Type F
Type C
300 300
250 250
200 200
150 150
100 100
50 50
0 0
0 5 10 15 20 0 5 10 15 20
magnitude of vertical displacement at the magnitude of vertical displacement at the
center(mm) center(mm)
(a) Regular honeycomb spokes (b )Auxetic honeycomb spokes
Fig. 10. Force–displacement curves of the NPTs when the cell wall thickness, t, of the spokes is 5 mm.

Fig. 11. (i) Force–deflection curve and (ii) deformed shapes with local stresses at vertical displacements of 10 and 20 mm of an NPT with the Type A honeycomb spokes.

Fig. 12. Honeycomb spokes designed based on Table 2.

2292 J. Ju et al. / Composite Structures 94 (2012) 2285–2295

is assigned. Loading and boundary conditions for the numerical
experiments will be described in detail in Section 4.3. As can be
seen in Fig. 7, the converged global mesh size of the tread is found
to be 0.005, which provides the total element number of the tread
– 1628.
Optimum element numbers for the spokes are also found by
checking reaction force and maximum local stress with varying
global mesh sizes for regular and auxetic honeycombs when the
obtained global mesh size (0.005 in ABAQUS) of the tread is as-
signed. As can be seen in Fig. 8, the global mesh size of the spokes
is found to be 0.0002, which provides the total element numbers of
the spokes ranging from 47196 to 65359.
The numbers of nodes and elements of the tread and the spokes
for each NPT are shown in Table 1. It should be noted that the
element numbers of the spokes are not necessarily the same for
each honeycomb spoke cell because each honeycomb spoke
geometry does not have the same vertical and inclined cell lengths.
For example, the Type F spokes have a longer vertical cell than the
other spokes have, which provides more element numbers for the
spokes. The element numbers of the tread are the same regardless
of the NPT spoke types (Table 1).
4.3. The same cell wall thickness based spoke design
Fig. 9 shows the deformed shapes of the honeycomb spokes
when a vertical displacement of 20 mm is applied at the center
of the hub. The reaction force at the center node is obtained from
ABAQUS when the vertical displacement loading is applied at the

Fig. 13. (i) Force–deflection curve and (ii) deformed shapes with local stresses at vertical displacements of 10 and 20 mm of an NPT with the Type B honeycomb spokes.

Fig. 14. (i) Force–deflection curve and (ii) deformed shapes with local stresses at vertical displacements of 10 and 20 mm of an NPT with the Type C honeycomb spokes.

Fig. 15. (i) Force–deflection curve and (ii) deformed shapes with local stresses at vertical displacements of 10 and 20 mm of an NPT with the Type D honeycomb spokes.
 J. Ju et al. / Composite Structures 94 (2012) 2285–2295 2293

Fig. 16. (i) Force–deflection curve and (ii) deformed shapes with local stresses at vertical displacements of 10 and 20 mm of an NPT with the Type E honeycomb spokes.

Table 3 The corresponding vertical force–deflection curves at the center

Dimension of cellular spokes and effective properties when designed to possess the of the hub are shown in Fig. 10 when a vertical displacement of
same reference load carrying capability. 20 mm is applied. Among the regular honeycomb spokes (Types
l (mm) h (mm) h (degree) t (mm) q⁄/qs Er/Es Eh/Es GrX/Es A–C), the Type A spokes show the highest reaction force with the
Type A 26.25 36.66 15.76 5.00 0.20 0.054 0.013 0.0016
vertical displacement due to its high effective modulus in the ra-
Type B 29.65 28.52 31.50 6.30 0.25 0.020 0.023 0.0062 dial direction; a load carrying capacity of 447 N at a vertical dis-
Type C 37.21 16.74 47.14 7.55 0.31 0.009 0.031 0.0378 placement of 20 mm. An increase in the cell angle induces the
Type D 26.25 50.93 15.76 4.85 0.23 0.049 0.012 0.0006 flexibility in the radial direction, which causes the honeycombs
Type E 29.65 59.43 31.50 6.74 0.36 0.025 0.028 0.0010
of Types B and C to have lower reaction forces; load carrying capac-
Type F 37.21 71.14 47.14 10.20 0.67 0.022 0.077 0.0020
ities of 340 N and 263 N at a vertical displacement of 20 mm for
NPTs with the Type B and the Type C honeycomb spokes, respec-
node of the hub center. The outer edge of the tread and the road tively. The auxetic honeycomb spokes show a similar force–
surface are set to have contact interaction when the vertical dis- deflection behavior as the regular counterparts; an increase in
placement loading is applied. While the vertical displacement magnitude of the cell angle induces the in-plane flexibility, result-
loading is applied at the hub center, the horizontal displacement ing in a lower reaction force. Additionally, an effective negative
of the bottom center on the tread is set to zero so that a deformed Poisson’s ratio of the auxetic honeycombs lowers the transverse
geometry can be maintained to be symmetric with respect to the reaction forces, resulting in a decrease in the overall reaction
plane perpendicular to the road surface; ux = 0 at x = 0 and y = 0 forces.
(Fig. 9). The honeycomb spokes with a lower magnitude of the cell an-
Each cell wall thickness of the spokes is designed to be 5 mm. As gle, e.g., NPTs with the Type A and the Type D spokes, cause a high-
mentioned in Section 3, the honeycomb spokes are modeled with a er chance of instability associated with elastic buckling of cells,
shear deformable beam element (B22 in ABAQUS). The hub and the whose nonlinear force–displacement is distinctive compared with
outer-face sheet are also modeled with the beam element. An eight the other spoke geometries (Figs. 9 and 10).
node plane stain element (CPE8R in ABAQUS) is assigned to the
tread. 4.4. Load carrying capacity based spoke design
The honeycomb spokes are subjected to bending at the contact
edges as well as the primary compression loading in the radial In this section, local stresses in the honeycomb spokes are
direction. Elastic buckling is observed in the Type A and Type D investigated when the NPTs are designed to have the same load
spokes. The dimensions of the honeycomb spokes and their carrying capability. Due to the nonlinear load carrying behavior
effective properties are shown in Table 2. as a function of vertical displacement, loading at a certain vertical

Fig. 17. (i) Force–deflection curve and (ii) deformed shapes with local stresses at vertical displacements of 10 and 20 mm of an NPT with the Type F honeycomb spokes.
2294 J. Ju et al. / Composite Structures 94 (2012) 2285–2295
Fig. 18. Total mass of NPTs with the honeycomb spokes for a lateral width of 100 mm.
displacement can be used as a reference value. Fig. 11 shows a ver-
tical force–deflection curve of an NPT with the Type A honeycomb
spokes when a vertical load is applied at the hub center.
As shown in Fig. 11, the effective force–deflection curve shows
the nonlinear behavior associated with combined nonlinear effects
of materials and geometries; primarily (i) hyperelastic material
behavior, and (ii) large deflection and buckling of cell walls of hon-
eycomb spokes, respectively. The deformed shapes of the cellular
spokes and maximum von Mises stresses at ‘1’ (a vertical displace-
ment of 10 mm) and ‘2’ (a vertical displacement of 20 mm) are
shown in Fig. 11; 3.5 MPa and 6.5 MPa at displacements of 10
and 20 mm, respectively.
In this study, the load carrying capability of an NPT with the
Type A honeycombs at a vertical displacement of 10 mm at the
hub center is set as the reference value; e.g., 294 N at a displace-
ment of 10 mm as shown in Fig. 11. It should be noted that the ref-
erence value does not necessarily mean the critical requirement for
the NPT design. In this study, the reference value is used for com-
parison with varying honeycomb spoke geometries.
Table 2 shows the dimensions and effective properties of the
honeycomb spokes when they are designed to have the reference
load carrying capacity of 294 N at a vertical deflection of 10 mm.
Fig. 12 shows the geometries of the honeycomb spokes based
on the dimensions in Table 2. Note that the cell wall thickness, t,
varies if NPTs have the same reference load carrying capacity (Ta-
ble 2). It should also be noted that the relative densities (q⁄/qs) of
auxetic honeycomb spokes (Types D, E, and F) have higher than
those of their counterparts.
Fig. 13 shows the force–deflection curve of an NPT with the
Type B spokes at the hub center when the tire is designed to pos-
sess the load carrying capacity of 294 N at a vertical displacement
of 10 mm. In order to meet the same load carrying capacity, the
Type B honeycomb spokes should have a cell wall thickness, t, of
6.3 mm. The initial slope of the force–deflection curve of an NPT
with the Type B spokes is lower than the slope with the Type A
spokes due to a lower Er/Es of the Type B spokes. The maximum
von Mises stresses are 2.2 MPa and 4.5 MPa at vertical displace-
ments of 10 mm and 20 mm, respectively. The lower local stresses
of the Type B spokes are favorable in designing fatigue resistant
spokes due to the lower cyclic stresses. More interestingly, the
local von Mises stresses of the NPT with the Type B spokes
(4.5 MPa) at a vertical displacement of 20 mm is lower than that
of an NPT with the Type A spokes (6.5 MPa), even though the mac-
roscopic load level of an NPT with the Type B spokes is higher than
the load level with the Type A spokes at the vertical displacement.
This might be caused by the local cell rotation effect.
The vertical force–deflection curve of an NPT with the Type C
spokes shows a fairly linear behavior compared to the former
NPT structures due to the in-plane uni-axial flexibility (Fig. 14).
The cell wall thickness, t, of the spokes is designed to be
7.55 mm to meet the reference load caring capacity. The increase
in the cell wall thickness leads an increase in the mass of the
spokes. A quantitative study on the mass of the NPTs will be dis-
cussed later in this section. Local stress levels of the Type C honey-
comb spokes further decrease, which turns out to be better for the
fatigue resistant design of the honeycomb spokes; the maximum
von Mises stresses are 1.9 MPa and 4.0 MPa at vertical displace-
ments of 10 mm and 20 mm, respectively.
Force–deflection curves of NPTs with the honeycomb spokes
having negative cell angles and their deformation are investigated.
Fig. 15 shows the force–deflection curve of an NPT with the Type D
honeycomb spokes whose cell wall thickness, t, is 4.85 mm associ-
ated with the reference load carrying capability design. A similar
macroscopic force–deflection behavior as an NPT with the Type A
spokes is noticed due to the similar effective moduli between the
Type A and the Type D spokes in the radial and the circumferential
directions. The maximum local stress levels of the Type D spokes
are also almost the same as those of the Type A spokes. Large cell
deformation of the Type D honeycomb spokes is also observed just
as it was in the Type A spokes, which might be caused by the low
cell wall thickness design leading to easy deformation of cell walls,
including buckling (Fig. 15).
A lower geometric nonlinear effect is observed with the Type E
spokes having a cell wall thinness, t, of 6.74 mm as shown in
Fig. 16. The corresponding maximum von Mises stresses are
2.3 MPa and 4.7 MPa at the global central displacements of 10
and 20 mm, respectively, which are lower than those with the
Type D spokes. The auxetic honeycomb spokes require more mass
to meet the reference load carrying capacity; for example, the
Type D and the Type E spokes have about 15% and 80% increased
 J. Ju et al. / Composite Structures 94 (2012) 2285–2295
mass, respectively, compared with that of the Type A spokes
(Table 3).
The Type F spokes have a higher t, resulting in an increase in
mass to meet the reference load carrying capacity; a 235% mass in-
crease compared to the Type A spokes (Table 2). The global force–
deflection curve of an NPT with the Type F spokes having a t of
10.2 mm is shown in Fig. 17. The local maximum von Mises stres-
ses in the Type F spokes further decrease; 1.7 MPa and 3.8 MPa at
‘1’ and ‘2’, respectively. A geometric nonlinear effect is rarely ob-
served compared to the former honeycomb spokes. This leads the
spokes to have the lower local stress values, which is good for fa-
tigue resistant honeycomb spokes.
The honeycomb spokes with a high cell angle magnitude
show low local stresses, which is good for the fatigue resistant
spoke design. The honeycomb spokes of Types C and F are good
for fatigue resistance. In terms of lower mass design, the Type C
spokes are thought to be good among the honeycomb spokes
investigated.
Fig. 18 shows the total mass of NPTs with the honeycomb
spokes when the lateral width is set to be 100 mm. As can be con-
firmed, an NPT with the Type F spokes has the highest mass, which
is not good for light weight design. The mass of an NPT with the
Type C spokes is 18% higher than that of Type A. In order to design
a tire for both low mass and high fatigue resistant honeycomb
spokes, a higher modulus elastomer as a base material with the
Type C spokes is preferable when designed to possess the same
load carrying capacity.
5. Conclusions
In this paper, a structural application of the flexible in-plane
properties of hexagonal honeycombs was suggested – the honey-
comb spokes of an NPT to replace the air of a pneumatic tire. Cel-
lular spoke geometries for an NPT were investigated with regular
and auxetic honeycomb spokes using the compliant cellular design
concept. The major findings are as follows:
 The ratio of the inclined cell length, l, to the overall cell height,
X1, is the key factor to determine the in-plane flexibility of hex-
agonal honeycombs under uni-axial loading.
 A high cell angle magnitude induces the high flexibility of the
honeycomb spokes, resulting in a lowering of the reaction force
of an NPT for a given vertical displacement loading.
 The honeycomb spokes with a higher cell angle magnitude
show lower local stresses, which is good for a fatigue resistant
spoke design, e.g., Types C and F. However, the honeycomb
spokes of Type C are better considering both fatigue resistance
and lower mass design.
 Experimental validation with prototypes of the NPTs is planned
for future work.
Acknowledgment
This work was supported by the 2010 Korea Aerospace Univer-
sity Faculty Research Grant.
References
[1] Gent AN, Walter JD. The pneumatic tire. Washington DC: National Highway
Traffic Safety Administration; 1985.
[2] Cozatt CP. Spring wheel. US patent, US 2502,908; 1924.
[3] Alfredo RV. Airless tire. US patent, US 3,329192; 1967.
[4] Kubica W, Schmidt O. Self-supporting motor vehicle tire. US patent, US
4169,494; 1979.
[5] Palinkas RL, Page GJ. Non-pneumatic tire with supporting and cushioning
members. US patent, US 4832,098; 1989.
2295
[6] Pajtas SR. Honeycomb non-pneumatic tire with a single web on one side. US
patent, US 4945,962; 1990.
[7] Panaroni VF. Airless tire with circumferential band reinforcement. US patent,
US 5086,815; 1992.
[8] Johnson TJ. Wheel assembly. US patent, US 5251,965; 1993.
[9] Kejha JB. Lightweight, safe and damage resistant tire having honeycomb core.
US patent, US 5685,926; 1997.
[10] Ohm TR. Resilient wheel structure. US patent, US 6615,885 B1; 2003.
[11] Laurent D, Delfino A, Hinc H. Non-pneumatic resilient tire. US patent, US
6640,859 B1; 2003.
[12] Rhyne T, Cron SM, Pompier JP. Compliant wheel. US patent, US 7013,939 B2;
2006.
[13] Rhyne T, Cron SM. Development of a non-pneumatic wheel. Tire Sci Technol
2006;34:150–69.
[14] Manesh A, Terchea M, Anderson B, Meliska B, Ceranski F. Tension-based non-
pneumatic tire. World Intellectual Property Organization, WO 2008/118983
A1; 2008.
[15] Kwark BM, Yoo EJ, Sung YH, Jang IG. Pattern design of non-pneumatic tire for
stiffness using topology optimization. In: Proceedings of the ASME
international design engineering technical conferences, DETC2009-86068,
San Diego, CA; 2009.
[16] Bitzer T. Honeycomb technology. London, UK: Charpman & Hall; 1997.
[17] Vinson JR. The behavior of sandwich structures of isotropic and composite
materials. Lancaster, PA: Technomic; 1999.
[18] Larsen UD, Sigmund O, Bouwstra S. Design and fabrication of compliant
micromechanisms and structures with negative Poisson’s ratio. J
Microelectromech Syst 1997;6(2):99–106.
[19] Olympio KR, Gandhi F. Zero-m cellular honeycomb flexible skins for one-
dimensional wing morphing. In: Proceedings of the 48th AIAA/ASME/ASCE/
AHS/ASC structures, structural dynamics, and materials conference,
AIAA2007-1735, Honolulu, Hawaii; 2007.
[20] Bubert E, Woods KS, Kothera CS, Wereley NM. Design and fabrication of a
passive 1-D morphing aircraft skin. In: Proceedings of the 49th AIAA/ASME/
ASCE/AHS/ASC structures, structural dynamics, and materials conference,
AIAA2008-2045, Schaumburg, IL; 2008.
[21] Spadoni A, Ruzzene M. Static aeroelastic response of chiral-core airfoils. J Intell
Mater Syst Struct 2007;18:1067–75.
[22] Heo H, Ju J, Kim DM, Jeon CS. Passive morphing airfoil with honeycombs. In:
Proceedings of the ASME international mechanical engineering congress and
exposition, IMECE2011-64350, Denver, CO; 2011.
[23] Abd El-Sayed FK, Jones R, Burgess IW. A theoretical approach to the
deformation of honeycomb based composite materials. Composites
1979;10(4):209–14.
[24] Gibson LJ, Ashby MF, Schajer GS, Robertson CI. The mechanics of two-
dimensional cellular materials. Proc Roy Soc A 1982;382:25–42.
[25] Masters IG, Evans KE. Models for the elastic deformation of honeycombs.
Compos Struct 1996;35:403–22.
[26] Bezazi A, Scarpa F, Remillat C. A novel centresymmetric honeycomb composite
structure. Compos Struct 2005;71:536–64.
[27] Balawi S, Abot JL. A refined model for the effective in-plane elastic moduli of
hexagonal honeycombs. Compos Struct 2008;84:147–58.
[28] Gonella S, Ruzzene M. Homogenization and equivalent in-plane properties of
two dimensional periodic lattices. Int J Solid Struct 2008;45:2897–915.
[29] Wang AJ, McDowell DL. In-plane stiffness and yield strength of periodic metal
honeycombs. Trans ASME J Eng Mater Technol 2004;126:137–56.
[30] Fan H, Jin F, Fang D. Uniaxial local buckling strengths of periodic lattice
composites. Mater Des 2009;30:4136–45.
[31] Ju J, Summers JD, Ziegert J, Fadel G. Compliant hexagonal meso-structures
having both high shear strength and high shear strain. In: Proceedings of the
ASME international design engineering technical conferences, DETC2010-
28672, Montreal, Quebec, Canada; 2010.
[32] Ju J, Summers JD. Compliant hexagonal periodic lattice structures having
both high shear strength and high shear strain. Mater Des 2011;32(2):
512–24.
[33] Berglind L, Ju J, Summers JD. Method to design honeycombs for a shear flexible
structure. SAE Int J Passenger Cars – Mech Syst 2010;3(1):588–97.
[34] Ju J, Ananthasayanam B, Summers JD, Joseph P. Design of cellular shear bands
of a non-pneumatic tire – investigation of contact pressure. SAE Int J Passenger
Cars – Mech Syst 2010;3(1):598–606.
[35] Ju J, Summers JD, Ziegert J, Fadel G. Design of honeycomb meta-materials for
high shear flexure. In: Proceedings of the ASME international design
engineering technical conferences, DETC2009-87730, San Diego, CA; 2009.
[36] Danielson KT, Noor AK, Green JS. Computational strategies for tire modeling
and analysis. Comput Struct 1996;61:673–93.
[37] Tonuk E, Unlusoy YS. Prediction of automobile tire cornering force
characteristics by finite element modeling and analysis. Comput Struct
2001;79:1219–32.
[38] Noor AK, Peters JM. Reduction technique for tire contact problems. Comput
Struct 1996;60:223–33.
[39] Kanyanta V, Ivankovic A. Mechanical characterization of polyurethane
elastomer for biomedical applications. J Mech Behav Biomed Mater 2010;3:
51–62.
[40] Pina-Hernandez C, Hernandez L, Flores-Velez LM, del Castillo LF, Dominguez O.
Processing and mechanical properties of natural rubber-ZnFe2O4
Nanocomposites. J Mater Eng Perform 2007;16:470–6.